AI Tool for Diabetes and Hypertension

IIITH researchers in collaboration with WISH Foundation are developing an automated web-based browser that will take critical healthcare decisions on behalf of primary healthcare workers. The project is sponsored by Google under its AI for Social Good program.

While the concept of telemedicine can be traced way back to the 1950s when it first evolved as a means of connecting patients residing in remote areas with medical professionals, its usage has surged manifold during the Covid-19 pandemic. With lockdowns, high infection rates and a physically overburdened healthcare system, telehealth has emerged as a safer means of interaction between patients and physicians. And when you add AI to this mix, it contributes to an overall better patient experience not just by eliminating the long waiting room times, but by providing accurate data-driven diagnosis and treatment.

Automated Decision-Making
IIITH researchers who have been engaged in the pursuit of applied healthcare research are currently developing an AI-powered web browser that can accurately assess patient risk based on prior medical history and the current line of treatment. The predictive tool will assist teleoperators or the primary healthcare providers in taking a decision of either continuing home-based care or flagging an emergency and consequently referring to a superior facility for specialised treatment.

Joint Venture
The initiative is a part of I-Hub-Data ((https://ihub-data.iiit.ac.in) which is a technology innovation hub established under the National Mission on Interdisciplinary Cyber Physical Systems (NM-ICPS) in the area of data-driven technologies. Speaking of the project’s initial focus, its principal investigator and the Academic Head of I-Hub-Data Prof. Deva Priyakumar says, “Non-communicable diseases such as diabetes and hypertension contribute to the bulk of the disease burden in the Indian population. In addition to this, an increase in the coexistence of these two diseases has been observed recently. There is also a greater emphasis on prevention rather than therapy, as well as a significant increase in the use of evidence and data to speed up innovation.” I-Hub’s goal to amplify research in broad data-driven technologies along with its dissemination and translation across the country ties in with Google Research India’s vision of AI with real-world impact. As per its AI for Social Good (AI4SG) website, they aim to support academic researchers, NGOs and other organisations in a variety of fields such as public healthcare. The data partner in this collaborative effort is the Wadhwani Initiative for Sustainable Healthcare (WISH) Foundation which is the flagship program of the Lords Education and Health Society (LEHS), an NGO that aims to make primary healthcare accessible and equitable to all Indians.

Data Collection
For any machine learning solution, the model ought to be trained over a large set of data. In this case, researchers are using the electronic health records deployed at various health centres as well as government data sets that LEHS has access to as part of various technical collaborations. The data had been most actively collected from the states of Rajasthan, Madhya Pradesh, Assam and Uttar Pradesh. The Primary Healthcare Centres from across Rural and Urban regions of these states process the data collected via Point of Care Devices through the ANMs of the centres.

Beyond Decision-Making
The long-term goal of the healthcare enterprise is to have a robust home-based care system in place. The idea is that it will not end once a decision or a referral (to a healthcare facility) is made. The team will plan on monitoring the patient’s wellbeing long after his or her visit to the primary healthcare centre. It may involve home visits by ANMs, recording of vital information such as BP and sending notifications related to nutrition and diet at regular intervals. According to Prof. Deva, the project is in line with the vision of the government of India and LEHS’ existing MoUs with various state governments make it highly likely to succeed in primary healthcare.

 

  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •